Predation exerts one of strongest selection pressures on prey populations, as failure to evade predators can result in mortality, physical harm or stress-related fitness costs (Kerfoot and Sih, 1987; Preisser et al., 2005). Successful evasion requires prey to detect predators reliably and assess the risk they pose. To achieve this, prey integrate information from a variety of sensory modalities, including visual (Coates, 1980; Gerlai, 1993; Wisenden and Harter, 2001), chemical (Kats and Dill, 1998; Bronmark and Hansson, 2000; Hettyey et al., 2015), acoustic (Blumstein et al., 2008; Radford et al., 2012; Hettena et al., 2014), and mechanical cues (Janssen, 2004; Yack, 2016).

Aquatic vertebrates mostly rely on chemical (Chivers and Smith, 1998; Wisenden, 2000; Brown, 2003; McCormick and Manassa, 2008) and visual cues (Helfman, 1989; Smith and Belk, 2001; Murphy and Pitcher, 2005; McCormick and Manassa, 2008) for predator detection. Recent research revealed that fish mostly use chemical cues to indicate the presence of a predator and use visual cues to modify their response to the predator (Hartman and Abrahams, 2000; Chivers et al., 2001). Visual cues are often considered to be part of passive sensing systems since they rely on external conditions that are not controllable by the organism. One exception to this are cases where fish improve their vision by means of controlled illumination (Nelson and MacIver, 2006). Controlled illumination was previously believed to be limited to bioluminescent organisms (Hellinger et al., 2017; Howland et al., 1992; Kenaley et al., 2014; Sutton, 2005). However, recent findings suggest that diurnal fish such as the triplefin, Tripterygion delaisi, are also capable of using active illumination. As is the case for many fishes, triplefins use the spherical lens that protrudes from their pupil to focus downwelling light on a bright patch located on the iris below the pupil. This is called an “ocular spark” and can be turned on or off at will (Michiels et al., 2018). It is proposed that they can use their ocular sparks to generate perceptible reflections in the retroreflective eyes of nearby prey or cryptic predators, and defined this process as ‘Diurnal Active Photolocation (DAP)’ (Bitton et al., 2019; Santon et al., 2020). For DAP to be effective, three conditions must be met:

At increasing depths, clear marine light fields undergo a fundamental shift: Illumination changes from direct downwelling to bluish, scattered light (Warrant and Johnsen, 2013). This change in both the angular distribution and spectral quality significantly reduces the efficacy of ocular-spark-based light redirection. Clear deep water, however, provides excellent conditions for red fluorescence to stand out (Neumeyer et al., 2002; Harant et al., 2016, 2018; Wilkins et al., 2016; Bitton et al., 2017). This is not the case in shallow water: Although there is more excitation light available here, any red fluorescence is flooded by the overwhelmingly strong red component of the shallow light spectrum.

Triplefins, along with many other small benthic fish species (2–5 cm), possess strong red fluorescent irises (Figure 1) (Anthes et al., 2016; Michiels et al., 2018). In T. delaisi, fluorescence peaks at 609 nm with a broad excitation spectrum peaking in the green range (530 nm) (Bitton et al., 2017). Triplefins regulate their fluorescence in response to ambient light conditions and show significantly stronger fluorescence in waters deeper than 20 meters. This, combined with the fact that they can perceive their own fluorescence, suggest that it serves a visual function (Michiels et al., 2008, 2018; Wucherer and Michiels, 2012, 2014; Kalb et al., 2015; Harant et al., 2016). Because of this we hypothesize that a red fluorescent iris can function as a light source for DAP in deeper water, e.g. below 20 m.

In this study, we test whether the red fluorescent iris of T. delaisi improves its ability to visually detect a nearby black scorpionfish (Scorpaena porcus), a cryptic sit-and-wait predator with strong retroreflective eyes (Schwab et al., 2002; Fritsch et al., 2017; Santon et al., 2018, 2020). To test this, we performed two experiments. First, we tested if red fluorescence affected the behavioural response of a triplefin to a nearby scorpionfish. To do so we placed different filters that blocked specific wavelengths between the triplefin and the scorpionfish to control whether the triplefin’s red fluorescence was allowed to contribute to the illumination of the scorpionfish or not. In the second, triplefins were confronted with a 3D-printed model of a scorpionfish with either a retroreflective or a non-retroreflective eye. A third group was tested with live scorpionfish as a positive control.

We tested whether the red light emitted by a triplefin’s fluorescent iris enhances its ability to detect a nearby scorpionfish (Experiment 1), and whether this can be explained by inducing and detecting fluorescence-induced eyeshine in the scorpionfish (Experiment 2).

Experiment 1 shows that bobbing, an indicator of seeing something “potentially dangerous” in triplefins (Santon et al., 2021), was particularly high in response to a scorpionfish relative to a stone, regardless of the FLUO treatment. This confirmed that bobbing in triplefins can be considered a reliable indicator of perceived danger, whereas a lack of it is associated with a safe visual scene (Smith, 1989; Graw and Manser, 2007; McCormick and Manassa, 2008; Calhoun et al., 2025; Geiger et al., 2025). Moreover, in the presence of a scorpionfish, bobbing was observed significantly more when fluorescence was “ON” relative to the “OFF” situation. This suggests that red fluorescent irises may indeed contribute to early detection of a scorpionfish.

However, these results cannot demonstrate that fluorescence was used to induce and detect retroreflective eyeshine in scorpionfish. Moreover, we did not check whether the scorpionfish moved more under the Fluo-ON than under the Fluo-OFF treatment. These issues were addressed during experiment 2. Here we showed that live scorpionfish did not move more in the Fluo-ON treatment and that there was no correlation between scorpionfish movement and triplefin bobbing frequency. Bobbing in triplefins is not exclusively associated with perceived threats, as triplefins also exhibited this behaviour when confronted with a stone. The comparatively low number of bobs observed in this context may be explained by motion parallax, a depth perception mechanism in which objects at varying distances appear to move at different speeds as the observer moves its head (Gibson et al., 1959; Kral, 2003; Bruckstein et al., 2005).

Experiment 2 confirmed a strong bobbing response toward live scorpionfish, with a significant difference between the fluorescence ON and OFF conditions, consistent with the results of Experiment 1. In contrast, responses to the retroreflective models were considerably weaker. Although the fluorescence ON/OFF comparison for these models showed only a subtle trend in the predicted direction, triplefins displayed more frequent bobbing toward retroreflective models with fluorescence ON than toward both non-retroreflective model treatments. This difference was not present when comparing the retroreflective model with fluorescence OFF against both non-retroreflective models. This suggests that models lacking retroreflective eyes are perceived as neutral objects, similar to stones, whereas retroreflective models elicit a response indicative of fluorescence-based retroreflection. The weak trend observed here might become more pronounced if the models were more lifelike (Brown and Warburton, 1997). Overall, these findings reinforce the idea that fish integrate multiple sensory modalities when detecting potential predators (Chivers and Smith, 1998; Blumstein et al., 2008; Radford et al., 2012; Hettena et al., 2014; Hettyey et al., 2015).

Research on active photolocation has mainly focused on two different light sources: redirected downwelling light (Bitton et al., 2019; Neiße et al., 2020; Santon et al., 2020) and bioluminescent organs (Howland et al., 1992; Hellinger et al., 2017) as the main light sources. Although it has been proposed that fluorescence could function as a light source for active photolocation (Jack, 2014), most of the studies on fluorescence focused on alternative function such as camouflage, photo-enhancement, intraspecific communication, prey attraction, photoprotection, or stress mitigation (Haddock and Dunn, 2015; Ben-Zvi et al., 2022; Poding et al., 2024). The use of red fluorescence for active photolocation can be expected to be limited to short-ranged interactions due to the low amounts of light involved. This aligns well with the observation that small benthic fishes in particular evolved red fluorescent irises, which may facilitate short-range visual predator detection (Anthes et al., 2016; Michiels et al., 2018).

Our findings highlight the potential for red fluorescence to serve as an alternative light source for diurnal active photolocation in blue-green environments at depth or in permanent shade, where bluish, scattered light is the dominant illuminant.